Abstract:
Previous research on interrupted task performance has focused heavily on factors that affect the resumption lag, the time it takes to resume from an interruption to the main task (Altmann & Trafton, 2002; Hodgetts & Jones, 2003,2006). More recently, several studies have found that the resumption lag tends to get shorter over repeated exposures - even after accounting for general task learning (Trafton, Altmann, Brock & Mintz, 2003; Cades, Boehm-Davis, Trafton & Monk, 2011). The purpose of this set of experiments was to evaluate the ability of the three main theories of interrupted task performance (Memory for Goals, Threaded Cognition, and Long Term Working Memory) to explain the reductions seen in resumption lag over repeated exposures. Experiment 1 examined the shape of the resumption lag distribution over repeated exposures as each theory suggested a different distribution. Findings suggested that the mechanism proposed by Long Term Working Memory was unlikely to be causing the reduction in resumption lag. Experiment 2 tested a specific mechanism postulated by the theory of Threaded Cognition (problem state). The results suggested that the reduction in resumption lag was not likely the result of the problem state mechanism. Experiment 3 examined the speeded encoding mechanism derived from Memory for Goals; the results indicated that improved ability to encode while performing the interruption task was not driving the reduction in resumption lag over repeated exposures to interruptions. Together, this set of experiments does not definitively support any single theory; however, the data do suggest which mechanisms are not likely to be driving the reductions seen in resumption lag over time, and suggest future experiments that may be able to determine the mechanisms underlying the reductions in resumption lag.
